Extracting relaxon velocities

[elliotperviz]

Dear community,

I have just started running some calculations with the Phoebe code, and have a question regarding the extraction of relaxon velocities.

Specifically, I am interested in extracting the relaxon mode velocities, for the purpose of calculating the modal contribution to kappa of the different relaxon modes (i.e. calculating the individual terms that enter the sum over relaxon modes in Eq. 13 in https://doi.org/10.1103/PhysRevX.6.041013).

As I understand (taking the same paper above as reference) I need to construct the product
V_alpha = < 0 | v | alpha >

From my tests, I see that the output file 'ph_relaxons_eigenvectors.hdf5' contains the relaxon eigenvectors (theta^{alpha}) and theta^{0}, which give two of the components required to calculate the relaxon velocities. What I don't have from the output are the phonon group velocities (v).

I also tested the phononDOS/phononBands tools, which do not output the group velocities either.

Of course, I can extract the necessary quantities from phonopy, but I would like to be sure that the same quantities are being used internally by phoebe. I would expect this to be the case if I am using the same force constants and same q-point mesh, but for whatever reason differences in implementation can lead to differences in the output.

So, the question is, whether it is sufficient to use phonon group velocities from phonopy? Otherwise, is it possible to output the phonon group velocities in phoebe, either as a relaxon output, or otherwise in the band/dos codes (in case I missed something in the documentation)?

I noticed that a similar question was asked once before (#220) but I am not sure if the feature was implemented or not.

Thanks!